The meme of innate class superiorities

by Leon Feinstein, Genomics Forum Bright Ideas Fellow

I am extremely grateful to the ESRC Genomics Forum in Edinburgh for the opportunity to spend a week there, away from the day job, reflecting on the question of “hereditarianism” and the relationship to it of Socioeconomic Status (SES) and child development, and to discuss this meme with experts. The Forum kindly hosted a workshop on March 6 2013, with a carefully invited group comprising psychologists, geneticists, social scientists, educationalists and representatives from the world of policy. The presentation built on recent helpful discussions at the Institute of Social and Economic Research (ISER) in Essex and at the Centre for the Analysis of Social Exclusion (CASE) at the LSE.

Download - The hereditarian view of social mobility presentation slides >>

I went to the workshop intending to learn how those in biological sciences have been thinking about the challenges of this issue. I motivated the discussion, as they say, with a focus on the technical issue of regression to the mean, which rather dominated proceedings. This has become something of a vexed issue for me, in relation to the graph that some know as “the Feinstein Graph.” I know it as Feinstein, 2003; Fig 2 but I am told I am stuck with the handle.

We didn’t really get into the issue of “hereditarianism” in quite the way I had expected but I arrived with a hope to learn something and was greatly informed by the experience.

I do not attempt a record of the event but I do want to thank participants for attending and their universally helpful contributions.

Please see attached a draft paper in relation to regression to the mean, which I managed to progress due to the discussion at the seminar. I think many participants wanted to have a wider discussion so this may be of minority interest.

As I say in the note, I do not at all accept that I have “misused statistics,” or that the graph "is simply wrong." The first is a glib slur and the second an odd statement because so much of the debate is about what the data mean.

I was further reminded at the Forum that work on this theme must address the changing reliability of the scores as children mature. This is complicated by the fact they are positions in an ordinal distribution of “ranks” in a weighted index of scores, rather than absolute performance on a single, specific test. People tend to forget this. It was helpfully suggested by the inestimable Tim Bates how further research to test the reliabilities of this sort of measure in a new sample might be carried out. I am hoping to discuss further. He said I should read Eysenck and Jensen.

Amongst many insightful points, Wendy Johnson emphasised particularly the importance of recognising that Fig 2, is really just a corollary of Fig 1. I strongly agree with this. I will pick this up in further work.

The fellowship at the Forum was a fantastic opportunity and I am especially grateful to the organisers.